Topic I
What Should Be the Underpinnings and Motivating Science and Applications Questions in a New Science of Hydrologic Extremes?

PRESENTATION

Upmanu Lall of Columbia University presented on the first, challenging task of outlining the questions scientists and practitioners need to be asking themselves in a “post-stationary” world. He noted that of all the many topics of interest in flood hydrology science, the past can be summarized as “the hydrologic engineer was in charge.” That is, the focus was primarily on a small subset of topics such as the design of structures, whether standards-based or risk-based, and insurance issues. Above all, the assumption was that of static risks. But this is changing with the acknowledgement that flood risk is conditional or dynamic, depending on watershed or land use conditions, climate conditions and the changing distribution of populations and assets. Standard practice in the past, for river flooding, is examining frequency and severity of floods as if these are invariant over a long enough period of record. Given the epochal or quasi-cyclical nature of climate variability that is organized over decadal and longer time scales, and flood records that are a few decades long, it is inevitable that flood frequency and severity will look very different as a transition across epochs is observed. This could be confused with anthropogenic changes in climate. A defensible and successful approach to mapping changing flood risk in time would need to recognize both systematic natural climate variations over decadal scales and anthropogenic changes in climate. The challenge is to develop a framework to encompass assessment of conditional (given observed climate conditions, as reflected in slowly varying ocean temperature fields) or dynamic (i.e., varying over time and indexed to climatic conditions) risk. Infrastructure operation and design could then use these more precise estimates of flood risk relevant for those conditions instead of reflecting the high uncertainty associated with the static risk estimate that lumps these very different conditions into a single flood risk measure that is invariant with time.

Lall then raised some “open questions,” as he termed them. The most prominent of these was whether we can do a better job of learning from major storms like Hurricane Katrina. A second was how we can generate a scientific basis for spatially explicit risk and exposure analysis, especially in the urban setting, and also over a large river basin. Finally, he wondered aloud whether we might develop ways to process information on short term vs. long term risk, incorporating uncertainty and discounting over time. This would have a bearing on decisions on infrastructure design as well as on financial risk management.

Lall had three additional concerns on the climate side. First, he said, for combined sea-level rise and wind surge we have not done a good job of developing detailed scenarios for the analysis of these risks for urban infrastructure and land use investments that are expected to last 30 to 50 years, or more. And



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 3
Topic I What Should Be the Underpinnings and Motivating Science and Applications Questions in a New Science of Hydrologic Extremes? PRESENTATION Upmanu Lall of Columbia University presented on the first, challenging task of outlining the ques- tions scientists and practitioners need to be asking themselves in a “post-stationary” world. He noted that of all the many topics of interest in flood hydrology science, the past can be summarized as “the hydro- logic engineer was in charge.” That is, the focus was primarily on a small subset of topics such as the design of structures, whether standards-based or risk-based, and insurance issues. Above all, the assump- tion was that of static risks. But this is changing with the acknowledgement that flood risk is conditional or dynamic, depending on watershed or land use conditions, climate conditions and the changing distribu- tion of populations and assets. Standard practice in the past, for river flooding, is examining frequency and severity of floods as if these are invariant over a long enough period of record. Given the epochal or quasi-cyclical nature of climate variability that is organized over decadal and longer time scales, and flood records that are a few decades long, it is inevitable that flood frequency and severity will look very different as a transition across epochs is observed. This could be confused with anthropogenic changes in climate. A defensible and successful approach to mapping changing flood risk in time would need to rec- ognize both systematic natural climate variations over decadal scales and anthropogenic changes in cli- mate. The challenge is to develop a framework to encompass assessment of conditional (given observed climate conditions, as reflected in slowly varying ocean temperature fields) or dynamic (i.e., varying over time and indexed to climatic conditions) risk. Infrastructure operation and design could then use these more precise estimates of flood risk relevant for those conditions instead of reflecting the high uncertainty associated with the static risk estimate that lumps these very different conditions into a single flood risk measure that is invariant with time. Lall then raised some “open questions,” as he termed them. The most prominent of these was whether we can do a better job of learning from major storms like Hurricane Katrina. A second was how we can generate a scientific basis for spatially explicit risk and exposure analysis, especially in the urban setting, and also over a large river basin. Finally, he wondered aloud whether we might develop ways to process information on short term vs. long term risk, incorporating uncertainty and discounting over time. This would have a bearing on decisions on infrastructure design as well as on financial risk management. Lall had three additional concerns on the climate side. First, he said, for combined sea-level rise and wind surge we have not done a good job of developing detailed scenarios for the analysis of these risks for urban infrastructure and land use investments that are expected to last 30 to 50 years, or more. And 3

OCR for page 3
4 Research and Applications Needs in Flood Hydrology Science what, he asked, are the prospects for developing scenarios at national scales for inundation/salt water in- trusion and its chronic demographic and ecological impacts? Second, he noted that we have scarcely be- gun to develop a monitoring, prediction and risk zone mapping capability for glacial lake outburst flood- ing. Third, can we predict inter-annual and longer persistent wet spells (and their hydrology) that lead to persistent lake flooding? Returning to river issues, Lall summarized what he called “dogmas”; that extreme rainfall plus ante- cedent land conditions equals an extreme river flood, that topography and channel network structure lead to scaling laws for flood extremes, and that various principles for regionalization of flood frequency re- lated to area or statistical homogeneity are available. All, he said, assume stationarity, and we have done very limited causal analysis of climate mechanisms and their use in prediction of static or dynamic risk. For example, extreme floods are usually related to moisture pulled in from the oceans, not local convec- tion. Recognizing that it is large scale moisture transport that leads to such floods could revolutionize flood hydrology, since the emphasis could shift from modeling the flood wave on the land surface to the moisture transport or flood wave as it progresses from the oceanic source to the watershed. To begin to answer this question, Lall stated, we have to look at the intersection of these systems with the ocean-atmosphere-land system, and we have not systematically done that. We need to know how much moisture is out there and how it is being moved in the various atmospheric layers. It is important to know how the moisture is organized and how it is moving, Lall stated, because about 90 percent of the moisture transport takes place in 10 percent of the atmosphere. We need to know how much is being moved in the lower and the higher levels. Soil moisture, he said, is recognized as an important antecedent condition for flooding and also for lending persistence to the prevailing atmospheric regime. However, even the persistence in the soil moisture is induced by initial widespread rainfall which usually has an oceanic source and whose occurrence is marked by a well defined organization of ocean-land temperature fields. These lead to clustered ocean locations where strong convection (evaporation) takes place, and the resulting moisture in the atmosphere is then moved in relatively narrow bands along reproducible path- ways, much like “atmospheric rivers” that then translate into landfall and heavy rainfall over a large area where conditions for the convergence of lower atmosphere flow are ideal, leading to flooding. Our current global climate models are not very helpful at this scale. Agreement among the models is often poor, there are issues with downscaling them, and they suffer error propagation issues, Lall said. Much of the large-scale atmospheric moisture transport that corresponds to extreme floods appears to re- late to lower-level jets and their intersection with upper-level troughs. These are problems that need to be worked out, he added. Even the fundamental prediction of these models that precipitation should be in- creasing with temperature can be called into question. Yet for all these problems, Lall is optimistic that there may be some climatic predictability of floods. He pulled together empirical evidence correlating sea surface temperature and floods in West Coast states, spoke of the potential importance of “atmospheric rivers” in concentrating precipitation in certain regions of the U.S., and finally examined some compos- ites and event analyses of extreme floods in the U.S., Brazil, and India. Finally, Lall closed the circle on his presentation with his view of the key questions that need atten- tion if assumptions of stationarity are limiting our advancement. Overall, he asked, given that there is evidence of inter-annual variability in flood incidence related to organized large-scale climate, how best can we develop a capacity for dynamic risk assessment, specifically with respect to frequency, spatial ex- tent, and duration? If—as he stated in his talk—global and regional climate models are not likely to di- rectly inform precipitation, and statistical downscaling is suspect, are there empirical pathways to achieve the same goals? Can we focus work on global and regional climate models toward atmospheric moisture pathways and mechanisms for flood generation? And how can we restate the climate change-related

OCR for page 3
Topic I 5 flood questions so we focus on mechanisms and causal chains to avoid error propagation in our models? With respect to the latter, he asked, how will local convection and stability change affect modest localized floods? How will spatio-temporal changes in ocean thermal content change the structure of low level jets, and tropical and extra tropical cyclones that then lead to enhanced flood potential? And how will regional flood-frequency estimates—static or dynamic—be improved using climate information, and be used to inform adaptation? PLENARY DISCUSSION A brief discussion, led by Dennis Lettenmaier of the University of Washington, emphasized the lack of integration of climate modeling and hydrologic modeling. One participant noted that beginning per- haps 30 years ago the paradigm was that by generating a scientific understanding of the climate system, this would help us understand precipitation, which would help us understand floods and their effects. This investment appears not to have been fruitful as scientists hoped, at least from the perspective of flood hydrologists. Three reasons for this were touched on by various participants. First, atmospheric models still have difficulty resolving crucial precipitation related phenomena, such as “atmospheric rivers” and low-level jets. Second, much of the research was oriented toward average weather conditions rather than the ex- tremes associated with flooding. Third, hydrologists have tended not to integrate atmospheric conditions into their analysis of, for example, streamflow trends from stream gauge records, and atmospheric scien- tists have tended to do the reverse. Reanalysis of both kinds of data together may yet yield useful results. Finally, several participants noted that while 30 years ago a meeting like this to consider linked at- mospheric and hydrologic processes would never have occurred, hydrologists will need to improve their communication with social scientists, economists, insurers, biogeographers, and others to effectively un- derstand and address the flood vulnerability issue. BREAKOUT SESSION REPORT Rapporteur Eric Wood summarized the discussion in the first breakout group. He first discussed the relationship between flood science and flood engineering. There was considerable discussion in the ses- sion of the apparent divide between “little” floods and “big” floods. He also noted that there are issues with the estimation of probable maximum precipitation and probable maximum floods, and many in the group seemed to think that the methods need to be updated. This would require updating the scientific basis of such estimates, including a comprehensive review of probable maximum precipitation methods. Wood noted that we still lack understanding with respect to the causes and behavior of truly large flood events. However, beyond this, how would a better scientific understanding find its way into design guidance? It was not clear to many in the session how to move forward to provide a vision for doing the necessary research and obtaining the funding for that research. It seems clear that there is an interagency interest in having such research done, and there was discussion of increasing interagency efforts in such research. However, interagency groups have very little funding at present for this kind of effort. Last, Wood summarized discussion of the beneficial aspects of floods—for ecosystems, for example. There is very little discussion of this in most guidance documents. How should such beneficial aspects— to the extent that they may be understood—be integrated into guidance as well as into design? Overall, he stated, there appear to be a great number of issues that the science and engineering community needs to pursue.